1 / 60

Expectation Propagation in Practice

Expectation Propagation in Practice. Tom Minka CMU Statistics Joint work with Yuan Qi and John Lafferty. Outline. EP algorithm Examples: Tracking a dynamic system Signal detection in fading channels Document modeling Boltzmann machines. Extensions to EP. Alternatives to moment-matching

Download Presentation

Expectation Propagation in Practice

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Expectation Propagation in Practice Tom Minka CMU Statistics Joint work with Yuan Qi and John Lafferty

  2. Outline • EP algorithm • Examples: • Tracking a dynamic system • Signal detection in fading channels • Document modeling • Boltzmann machines

  3. Extensions to EP • Alternatives to moment-matching • Factors raised to powers • Skipping factors

  4. EP in a nutshell • Approximate a function by a simpler one: • Where each lives in a parametric, exponential family (e.g. Gaussian) • Factors can be conditional distributions in a Bayesian network

  5. EP algorithm • Iterate the fixed-point equations: • specifies where the approximation needs to be good • Coordinated local approximations where

  6. (Loopy) Belief propagation • Specialize to factorized approximations: • Minimize KL-divergence = match marginals of (partially factorized) and (fully factorized) • “send messages” “messages”

  7. EP versus BP • EP approximation can be in a restricted family, e.g. Gaussian • EP approximation does not have to be factorized • EP applies to many more problems • e.g. mixture of discrete/continuous variables

  8. EP versus Monte Carlo • Monte Carlo is general but expensive • A sledgehammer • EP exploits underlying simplicity of the problem (if it exists) • Monte Carlo is still needed for complex problems (e.g. large isolated peaks) • Trick is to know what problem you have

  9. Example: Tracking Guess the position of an object given noisy measurements Object

  10. Bayesian network e.g. (random walk) want distribution of x’s given y’s

  11. Terminology • Filtering: posterior for last state only • Smoothing: posterior for middle states • On-line: old data is discarded (fixed memory) • Off-line: old data is re-used (unbounded memory)

  12. Kalman filtering / Belief propagation • Prediction: • Measurement: • Smoothing:

  13. Approximation Factorized and Gaussian in x

  14. Approximation = (forward msg)(observation)(backward msg) EP equations are exactly the prediction, measurement, and smoothing equations for the Kalman filter - but only preserve first and second moments Consider case of linear dynamics…

  15. EP in dynamic systems • Loop t = 1, …, T (filtering) • Prediction step • Approximate measurement step • Loop t = T, …, 1 (smoothing) • Smoothing step • Divide out the approximate measurement • Re-approximate the measurement • Loop t = 1, …, T (re-filtering) • Prediction and measurement using previous approx

  16. Generalization • Instead of matching moments, can use any method for approximate filtering • E.g. Extended Kalman filter, statistical linearization, unscented filter, etc. • All can be interpreted as finding linear/Gaussian approx to original terms

  17. Interpreting EP • After more information is available, re-approximate individual terms for better results • Optimal filtering is no longer on-line

  18. Example: Poisson tracking • is an integer valued Poisson variate with mean

  19. Poisson tracking model

  20. Approximate measurement step • is not Gaussian • Moments of x not analytic • Two approaches: • Gauss-Hermite quadrature for moments • Statistical linearization instead of moment-matching • Both work well

  21. Posterior for the last state

  22. EP for signal detection • Wireless communication problem • Transmitted signal = • vary to encode each symbol • In complex numbers: Im Re

  23. Binary symbols, Gaussian noise • Symbols are 1 and –1 (in complex plane) • Received signal = • Recovered • Optimal detection is easy

  24. Fading channel • Channel systematically changes amplitude and phase: • changes over time

  25. Differential detection • Use last measurement to estimate state • Binary symbols only • No smoothing of state = noisy

  26. Bayesian network Symbols can also be correlated (e.g. error-correcting code) Dynamics are learned from training data (all 1’s)

  27. On-line implementation • Iterate over the last measurements • Previous measurements act as prior • Results comparable to particle filtering, but much faster

  28. Document modeling • Want to classify documents by semantic content • Word order generally found to be irrelevant • Word choice is what matters • Model each document as a bag of words • Reduces to modeling correlations between word probabilities

  29. Generative aspect model (Hofmann 1999; Blei, Ng, & Jordan 2001) Each document mixes aspects in different proportions Aspect 1 Aspect 2

  30. Generative aspect model Aspect 1 Aspect 2 Multinomial sampling Document

  31. Two tasks Inference: • Given aspects and document i, what is (posterior for) ? Learning: • Given some documents, what are (maximum likelihood) aspects?

  32. Approximation • Likelihood is composed of terms of form • Want Dirichlet approximation:

  33. EP with powers • These terms seem too complicated for EP • Can match moments if , but not for large • Solution: match moments of one occurrence at a time • Redefine what are the “terms”

  34. EP with powers • Moment match: • Context function: all but one occurrence • Fixed point equations for

  35. EP with skipping • Context fcn might not be a proper density • Solution: “skip” this term • (keep old approximation) • In later iterations, context becomes proper

  36. Another problem • Minimizing KL-divergence of Dirichlet is expensive • Requires iteration • Match (mean,variance) instead • Closed-form

  37. One term VB = Variational Bayes (Blei et al)

  38. Ten word document

  39. General behavior • For long documents, VB recovers correct mean, but not correct variance of • Disastrous for learning • No Occam factor • Gets worse with more documents • No asymptotic salvation • EP gets correct variance, learns properly

  40. Learning in probability simplex 100 docs, Length 10

  41. Learning in probability simplex 10 docs, Length 10

  42. Learning in probability simplex 10 docs, Length 10

  43. Learning in probability simplex 10 docs, Length 10

  44. Boltzmann machines Joint distribution is product of pair potentials: Want to approximate by a simpler distribution

  45. Approximations EP BP

  46. Approximating an edge by a tree Each potential in p is projected onto the tree-structure of q Correlations are not lost, but projected onto the tree

More Related